simulation-digital-twin-production-optimization

Minimize time wastage through simulation

Written by Vincent Béchard on 2023-06-20

Typical production planning process

How do you assess the quality and feasibility of production planning? Are operations still taking place as planned? Is there any unanticipated loss of time? If this is not always the case, then operations simulation is a tool to use during production planning. This allows activities to be scheduled to minimize wasted time and other unproductive waits, while allowing sufficient time to absorb process variation.

The typical weekly production cycle, for example in a food or beverage factory, is: obtaining the list of production orders, planning and execution of production. Production orders vary from week to week, so planning must be reviewed each week.

When planning, the complexity of the exercise arises from the many combinations and constraints to consider, for example: different products, different packaging, equipment characteristics, product-product and product-equipment compatibilities, sequences and batch sizes, etc. During execution, the difficulty of the exercise arises from the realities of the production floor: equipment reliability, product changeover times, various expectations, natural variability in flow rates and cycle times, etc.

At the end of the week: disaster, the plan did not go as planned! Overtime was required, orders were delivered late to customers, production teams blamed the unrealistic planning received, and planning teams blamed the efficiency of operations that was compromised. below expectations... What if it were possible to plan by anticipating how operations will take place?

Simulate for smarter scheduling

It is possible to capture the know-how of operations specialists in order to anticipate how a production schedule will be executed. One possible technique is based on discrete events simulation. It is an approach for modelling material flows (parts, bulk materials, flow rates), tasks to be carried out, production schedules and decision rules. Such models make it possible to fully understand the interactions between the elements of a production system and to estimate productivity. The use of random numbers is the key ingredient: this makes it possible to mimic the variability of processes, and therefore, to assess the risk of error and the precision of predictions.

Building a discrete events simulation model, when done according to the state-of-the-art techniques, requires involving the people on the production floor: it is the operators who are best able to provide the details and subtleties to model. Additionally, variability in actual performance and historically observed efficiencies are taken into consideration. So this type model is literally the production floor represented in a computer, in a convenient and fast format. It's a digital twin! It is possible to test many scenarios in order to find the most favorable… such as the realistic and optimal production schedule!

Minimize time wastage with simulation

Back with the weekly planning task: now that it is possible to anticipate the production response for a given schedule, planners can now adjust sequences, batch sizes and equipment assignments in order to:

  • Reduce waiting times
  • Reduce the number of product changeovers
  • Insert maintenance and inspections at the appropriate times
  • Adequately balance the workload between equipment
  • Ensure lots are completed on time for customers

Since production duration estimates are obtained by considering the numerous operating constraints as well as the various sources of variation and disruption, the probability that the weekly plan will be executed as planned is greatly increased.

And there are some side effects of using such simulations: the duration of the planning exercise is reduced and involves fewer people, planning methods are automatically standardized, it is easier to transfer knowledge -make use of other resources, the work climate is improved, and so on.

How much does it cost to build a simulation? How much does time wastage and their consequences cost?

Illustration with a dairy plant

A dairy product pasteurization and packaging plant may have to manage more than 100 formulations and more than 50 packaging formats, or more than 200 different products to be scheduled each week. Some of these products have incompatible colors, flavors, acidity and shelf life, or even allergens. The current normative framework imposes cleaning and sterilization frequencies. About 2/3 of the production must be made towards the end of each week to meet consumption habits. What a headache!

A model was developed using the Flexsim simulation software. This model considers elements such as: pasteurization modes and flow rates, quasi-artificial intelligence for the assignment of storage tanks, filling rates and efficiencies of each filler according to each product and each packaging format, and the rules governing cleaning and sterilization. The model interprets the production schedule and ensures that all batches are executed on time taking into account the elements mentioned above.

scheduled lots completed on time is evaluated for each filler on each day of the week. A schedule produced “normally” was evaluated by simulation (table on the left). Operators know that, for Thursdays and Fridays (days 5 and 6), schedules are generally unachievable. The table on the left paints exactly this portrait. After about an hour of working with the simulation, a planner produced a schedule whose performance appears in the table to the right. Without the simulation tool, the same exercise would have required 5 to 6 hours for a small team that did not have the computing power of the simulator.

simulation-aided-scheduling

Is this new schedule credible? The operators who performed it reported very few deviations and all batches were actually produced on time. The sum of all the waiting time of all the fillers (lack of product) amounted to 254 hours for the “normal” schedule, while for the same workload, the improved version of the schedule only included 60 hours of waiting. A 76% reduction in wasted time! And that's just operating time lost, it doesn't take into account the time saved during the planning process... Planners and managers have more time to devote to continuous improvement and increasing performance. customer satisfaction.

Do we think we can reduce cleaning and sterilization times by changing equipment allocation priorities? A few simulation checks confirm the best choices, and tests during the production week confirm the impacts. A planning nightmare occurs when new products are put into production for the first time by the marketing team... Never mind, with such a simulation potential impacts are anticipated and reduced before they present themselves.

A large promotional volume for a large supermarket banner is required? The simulation made it possible to quickly answer “yes, we can do it” with a supporting schedule. With 194 hours available to produce instead of waiting... Owner of a network of dairy plants wanting to reduce delivery costs by distributing volumes more efficiently between plants? About five hours of simulation work were spent finding the new, more economical distribution!

Read the full article here.

In conclusion

In summary, weekly production scheduling is essential to the operation of a plant. It is effective and realistic only if the details, complexities, constraints and subtleties of the operations are taken into consideration. To do this, discrete events simulation is an ideal modelling approach. Planners with such a simulation are able to adequately adjust schedules in order to reduce operational stress: fewer unforeseen events, more room for maneuver, balanced workloads, etc. Result: less wasted time, more production time.

Moreover, in a context aiming to approach a Lean production system, simulation is a tool making it possible to reduce several wastes both at the production and planning levels:

  • Production waste: waiting, workload imbalance, delivery time, unnecessary product changes, etc.
  • Planning waste: errors and rework, overprocessing, frustration
Finally, simulation makes it possible to introduce people to the impact of variability on performance and it also makes it possible to understand and evaluate the benefits of reducing the variability of certain elements critical to planning performance .

Want to learn more?

At Différence, our core expertise is centered on statistic and data science, Lean applications and operational excellence, and... simulation! We can train, coach and help practitioners to learn how to use simulation! Don’t hesitate to ask for more information by contacting us.